Dengue virus (DENV) infections are an emerging problem, causing disease in tropical and subtropical countries, including the U.S, with recent outbreaks in Texas, Hawaii and Florida. An increase in the average age of DENV related hospitalizations has been globally recognized with adult disease more commonly recognized, but prospective clinical studies that characterize dengue illness in both children and adults are lacking. Dengue hemorrhagic fever (DHF), a more severe clinical manifestation of DENV infection, has been classified by the WHO into four grades of severity with minimal criteria of capillary leakage and thrombocytopenia. The WHO released revised guidelines for the diagnosis of dengue and severe dengue in 2009. Severe dengue comprises not only plasma leakage and bleeding, but also end organ failure, which is a relatively rare event. In addition, several 'warning signs', such as abdominal pain and nausea, are highly non-specific. A major dilemma in the management of dengue illness is the difficulty in early recognition of those individuals who will go on to severe dengue and/or shock. We have previously developed a classification and regression tree algorithm for identifying such individuals at risk for severe illness. Frequent blood tests (white blood cell counts, hematocrit, liver function studies) are needed to identify severe disease, which is not practical in resource poor settings. Non-invasive monitoring which might capture individuals with impending complications and would help to guide therapy would greatly enhance the current management of DENV illness. We propose to address these issues with the following Specific Aims: I.To determine the clinical features which characterize severe disease in hospitalized children and adults with suspected dengue - characterize early and late disease in adults vs. children, evaluate old and new WHO criteria, compare immunologic and virologic correlates to severe disease in adults vs. children II.To develop and validate an algorithm which will enable physicians to predict which children or adults will develop significant plasma leakage and/or shock and to define those who will not. We will utilize novel noninvasive monitoring (near infrared spectroscopy and arterial wave forms) to accomplish these goals.